Urgent.News

What's breaking now, across thousands of outlets.

More in Tech

Are your model's probabilities calibrated? Expected calibration error in NumPy

There was a comment on Hacker News this week saying LLM token probabilities are badly calibrated for making decisions. I think that's right, and it's also something you can measure yourself in about…

  • Expected calibration error (ECE) measures calibration of model probabilities
  • NumPy library used to calculate ECE in Python code snippet
  • Perfectly calibrated model has lowest ECE and Brier score among four tested

Using NLTP with MQTT: Healing 35% Wireless Packet Loss on ESP-NOW & UDP Before Your Broker

If you have ever deployed outdoor sensor nodes over noisy 2.4 GHz radio links, ESP-NOW, or rural 4G CGNAT modems, you have probably hit the First-Mile Wall : MQTT over TCP is great inside the cloud…

  • NLTP protocol developed to address wireless packet loss in ESP-NOW and UDP
  • NLTP improves packet loss recovery by up to 35% before reaching MQTT broker
  • NLTP uses compact header, LZ4 compression, AEAD, and Reed-Solomon error correction

The Drift Problem That Breaks GitOps at Scale

The hard part of GitOps at scale is not the sync engine. It is drift: everything that changes your cluster after Argo CD says it is in sync. If your first post on why LinkedIn outgrew Argo CD covered when the engine breaks, this covers the quieter failure -- the one that makes your Git repo lie to you.

More from Saturday 14 November →